In this paper an approach for VoiceXML web application modeling is presented with which a voice application can be modeled as a call flow diagram that can be further transferred to be deployable Java code by an automa...
详细信息
In this paper an approach for VoiceXML web application modeling is presented with which a voice application can be modeled as a call flow diagram that can be further transferred to be deployable Java code by an automated code generator implemented in the same study. This paper focuses on the modeling work while the brief discussion on code generation is given as well.
Class Language Model(LM)has been proved to be usefulin the case of sparse training *** traditionally,it hasto be combined with word LM to get attractiveperformance,because the existing classification criteriaonly invo...
详细信息
Class Language Model(LM)has been proved to be usefulin the case of sparse training *** traditionally,it hasto be combined with word LM to get attractiveperformance,because the existing classification criteriaonly involve language context information or *** ovcrcome this,our paper presents a newmethod to cluster the words that have different soundingand similar language context into one *** that the class LM based on the proposed methodachieves good performance without combining with a wordLM.
Document Object Model(DOM)is widely used fordynamic description of eXtensible Markup Language(XML)document and its *** is anXML instance targeting voice interaction for enterpriselevel telephony applications.A design ...
详细信息
Document Object Model(DOM)is widely used fordynamic description of eXtensible Markup Language(XML)document and its *** is anXML instance targeting voice interaction for enterpriselevel telephony applications.A design on DOM extensionfor VoiceXML is thus introduced in this paper to providea standard way on VoiceXML manipulation and build anexposing mechanism for VoiceXML *** module and event module are presented indetail.
A voice conversion algorithm based on acousticfeature transformation is *** synthsis is widely used *** such asystem,voice conversion is much more challengingthan a paramter bases synthesizer,such as *** main idea is ...
详细信息
A voice conversion algorithm based on acousticfeature transformation is *** synthsis is widely used *** such asystem,voice conversion is much more challengingthan a paramter bases synthesizer,such as *** main idea is to constructtransformations of the acoustic features between twospeakers with Maximum Likelihood Linear Regression(MLLR)based on the acoustic feature domainconcatenative speech synthesis,Some results of aseries of experiments which are based on IBMtrainablc specch synthesis system,IBM cepstralreconstruction and LSP reconstruction algorithms arepresented.
Capitalizing on the short-timestationarity of speech signal, an estimation ofthe background noise parameters in noisyspeech signals is derived. The novel approach,requiring no active/silent frame detection tothe noisy...
详细信息
Capitalizing on the short-timestationarity of speech signal, an estimation ofthe background noise parameters in noisyspeech signals is derived. The novel approach,requiring no active/silent frame detection tothe noisy speech, can be computed effieientlyand give real-time estimation of the *** results can be achieved even if thebackground noise has slowly time varyingfeature, so the speech enhancement effect *** Terms—speech enhancement, noiseestimation, short-time spectral amplitude,spectral subtraction estimator
Word Acoustic Distance (WAD) measurement could help improving the performance of speech-enabled navigation or transaction applications. This paper presents a novel model-driven method for the WAD measurement between a...
详细信息
Word Acoustic Distance (WAD) measurement could help improving the performance of speech-enabled navigation or transaction applications. This paper presents a novel model-driven method for the WAD measurement between any two words. Enhanced n-best list generation based on the present approach was presented for illustration. An evaluation on WAD was conducted in a real-life Name Dialer situation and achieved a satisfying recalling rate of 94.8% with three thousand entries.
Dealing with polyphones is an important part of Chinesetext-to-speech *** the pronunciation of aChinese character is directly related to the meaning of it,an algorithm based on semantic calculation using How-Netis int...
详细信息
Dealing with polyphones is an important part of Chinesetext-to-speech *** the pronunciation of aChinese character is directly related to the meaning of it,an algorithm based on semantic calculation using How-Netis introduced to determine the pronunciations of thepolyphones in new words,which hasn’t appeared in thepolyphone list,the polyphone knowledge base or *** experiment results prove it can do goodperformance.
The homogeneous Hidden Markov Model (HMM) (or automatic speech recognition has been widely used today. But some significant defects of this method limit its performance and praclical applications. One of these defects...
详细信息
The homogeneous Hidden Markov Model (HMM) (or automatic speech recognition has been widely used today. But some significant defects of this method limit its performance and praclical applications. One of these defects is that the stability of duration distribution of the speech states which is verified by experiments[1] is not correctly considered in the model In this paper, a duration distribution based inhomogoneous. HMM(DDBHMM) recognition algorithm is introduced. A speaker-independent isolated-word Chinese speech recognition experiment is done and shows that DDBHMM reduces the error rale by about 20%.
暂无评论